Warning message for invalid key in configuration file #1498
Conversation
This commit fixes issue #1497.
…incubator-kvrocks into warning_message_for_invalid_key
gloof11
changed the title
|
Hi @gloof11 can you write a cpp unit test? (probably in https://github.com/apache/incubator-kvrocks/blob/unstable/tests/cppunit/config_test.cc) |
|
Thanks for your contribution! The code need to be formatted to pass the CI, refer to https://kvrocks.apache.org/community/contributing#code-style. |
|
It seems the code is still not formatted properly. And I think you need to give a reason why you do not just check |
…. Currently there is no clang-tidy for M1 Macs so here goes nothing
|
Just made a commit. I read up on how maps actually work in Cpp. In the refactor, if the key that was passed to "iter" was not a valid key via "find()", then it would be equivalent to "past-the-end". It simply checks if both values are considered "past-the-end", and throws an error message if that's the case. Also, I'm not at all familiar with writing unit tests for C++. Sorry :( |
|
I don't know whether it needs a test because we can't get a log message in th test. I think the code is obvious. |
|
@infdahai Yes, I also think that a test, in this case, is too fanatic. |
|
You need to revert the change of
|
|
Hi @gloof11 , any progress or problem here? |
|
@PragmaTwice You slashed the clang linter with the long line :) |
For the namspace/token pair, it shouldn't go through the configuration key check process, but we forgot to return when it's namespace key. And it will print an invalid configuration key warning message after apache#1498 which may confuse users.
This is a fix that I was able to come up with!
It's hacky, but it works.